AI Incidents
Real-world cases where AI systems caused harm, classified by risk domain and entity. Data sourced from 
AIID and classified by 
MIT AI Risk Initiative. CC BY-SA 4.0
1,751 incidents
Activists Allege NYPD's Application of Facial Recognition Interfered with Right to Protest
Black Lives Matter activists alleged being targeted for arrest by New York Police using facial recognition, interfering with their right to protest.
OpenAI's GPT-3 Associated Muslims with Violence
Users and researchers revealed generative AI GPT-3 associating Muslims to violence in prompts, resulting in disturbingly racist and explicit outputs such as casting Muslim actor as a terrorist.
Bug in Instagram’s “Related Hashtags” Algorithm Allegedly Caused Disproportionate Treatment of Political Hashtags
A bug was reported by Instagram’s spokesperson to have prevented an algorithm from populating related hashtags for thousands of hashtags, resulting in an allege preferential treatment for some politically partisan hashtags.
Colorado Police’s Automated License Plate Reader (ALPR) Matched a Family’s Minivan’s Plate to That of a Stolen Vehicle Allegedly, Resulting in Detainment at Gunpoint
An automated plate reader reportedly matched a license plate information, but of a family’s minivan and an alleged motorcycle in Montana that was reportedly stolen earlier in the year, resulting in them and their children being held at gunpoint and detained in handcuffs by multiple Aurora police officers.
Genderify’s AI to Predict a Person’s Gender Revealed by Free API Users to Exhibit Bias
A company's AI predicting a person's gender based on their name, email address, or username was reported by its users to show biased and inaccurate results.
Facebook Provided Offensive Translation for King of Thailand's Birthday Ceremony
Facebook's Thai-English translation gave an inappropriate mistranslation on Thai PBS's Facebook live broadcast of the King of Thailand’s candle-lighting birthday ceremony.
Thoughts App Allegedly Created Toxic Tweets
Tweets created by Thoughts, a tweet generation app that leverages OpenAI’s GPT-3, allegedly exhibited toxicity when given prompts related to minority groups.
Deepfake Images of Australian Teacher Hannah Grundy and 25 Others Created and Circulated Online by Former Friend
Australian teacher Hannah Grundy discovered her face had been superimposed onto pornographic images using AI deepfake technology, which were shared online alongside personal details. A former trusted friend, Andrew Thomas Hayler, was found responsible, targeting 26 women, including Grundy and close friends, over two years. Hayler pleaded guilty to multiple charges and was sentenced to nine years in prison.
International Baccalaureate’s 2020 Calculated-Grades Model Reportedly Lowered Students' Scores and College Prospects
After Covid-19 canceled May 2020 exams, the Geneva-based International Baccalaureate reportedly used a calculated-grades model based on coursework, predicted grades, and historical school results to estimate final marks. Students, parents, and teachers reportedly said the opaque model produced unexpectedly low scores, threatening university admissions, scholarships, and study plans across multiple countries.
NarxCare’s Risk Score Model Allegedly Lacked Validation and Trained on Data with High Risk of Bias
NarxCare's overdose risk algorithm, lacking peer-reviewed validation, uses sensitive data like doctor visits, prescriptions, and possibly genetic information, leading to significant biases against women and Black patients. Factors like sexual abuse and criminal records exacerbate stigmas and disparities, often resulting in unjust denial of necessary pain medication. The newly approved AvertD genetic test shares similar issues, further complicating and potentially harming medical treatment decisions.
Facebook’s Political Ad Detection Reportedly Showed High and Geographically Uneven Error Rates
The performance of Facebook’s political ad detection was revealed by researchers to be imprecise, uneven across countries in errors, and inadequate for preventing systematic violations of political advertising policies.
WeChat Pay's Facial Recognition Security Evaded by Scammers Using Victims’ Social Media Content
In China, fraudsters bypassed facial-recognition security for online financial transactions on WeChat Pay by crafting identity-verification GIFs of victims from their selfies on WeChat Moments, a social media platform.
Reported AI-Assisted Influence Campaigns by GoLaxy Allegedly Targeting Hong Kong and Taiwan Political Discourse
Reportedly leaked documents reviewed by researchers and media allege that Chinese firm GoLaxy used its "GoPro" AI system in multiple influence campaigns. In 2020, it reportedly tracked 180,000 Twitter accounts to counter opposition to Hong Kong's National Security Law; in 2022–2023, it amplified praise for China's Zero Covid policy using purportedly fake Facebook accounts; and in late 2023, allegedly it advised on messaging to weaken Taiwan's Democratic Progressive Party.
YouTube's AI Mistakenly Banned Chess Channel over Chess Language Misinterpretation
YouTube's AI-powered hate speech detection system falsely flagged chess content and banned chess creators allegedly due to its misinterpretation of strategy language such as "black," "white," and "attack" as harmful and dangerous.
Facebook's AI Put "Primates" Label on Video Featuring Black Men
Facebook's AI mislabeled video featuring Black men as a video about "primates," resulting in an offensive prompt message for users who watched the video.